The ORTDM occupies the eastern coastal portion of the Eastern Cape province, South Africa. The district lies along the coastline of the Indian Ocean stretching for up to 160 km (Morgan 2017). The district extends over a geographical area of 15946.84 km² and incorporates five local municipalities, referred to by Figure 1 (Morgan 2017). ORTDM lies between the coordinates of 32°46’31”S and 21°23’29”E (Mlanjeni 2014). OR Tambo is classified as a Category C2 municipality, indicating a largely rural character and low urbanisation rate. In addition to agriculture, the other economic sectors are community services, trade, finance, transport, manufacturing and construction (Null 2018). Its suitable terrain and many river valleys provide irrigable land, abundant water resources, large tracts of grazing land, suitable pasture species for stock grazing, and a large number of stock owned by rural communities (Davies 2014). The district has the richest natural resources and the most fertile soils and favourable climatic conditions. Agricultural practices are intense although climate dependent (Thomas 2014). These have diverse vegetation types ranging from grasslands, thicket, forests and bushveld (Munn 2018). Mlanjeni (2014) notes that drought incidents negatively impact agricultural production of the OR Tambo district and contribute to food insecurity. The OR Tambo district receives plenty of rainfall and increased humidity during summer. Winters are colder especially in inland areas. The district’s climatic conditions are moderate to high rainfall areas, mainly along its sub-tropical coast and in pockets of mountainous areas (Narloch & Bangalore 2018). The climatic conditions of OR Tambo district have warm, temperate, predominantly frost-free conditions. The OR Tambo district enjoys a high level of annual sunshine, and in summer, temperatures range from 16 ºC to 28 ºC whilst winter temperatures range from 7 ºC to 20 ºC. Winter months fall between April and August whilst summer temperatures are usually highest between November and April (Slattery 1998). The people in the OR Tambo district enjoy four seasons of the year namely, summer, winter, spring and autumn and they are characterised by different weather conditions ranging from hot, to cool, mild, windy and cold conditions.

Data used on this study were obtained at South African Weather Services (SAWS) in Pretoria. The study employed daily, monthly and annual precipitation data gathered from the SAWS for the period 1998 to 2018 and recorded the data using both automatic and manual weather stations located across ORTDM. There are a number of weather stations in the ORTDM. Before 1979, the municipality had 17 operating weather stations to assist in weather recording and forecasting. Advances in technology led to the addition of three automatic weather stations in ORTDM. The automatic weather stations are located in Port St Johns and King Sabatha Dalindyebo municipalities, see Figure 1. According to Mlanjeni (2014), manual weather stations are not advanced because they had a majority of disadvantages such as missing data and lacked coherence. Advances in technology and the introduction of automatic weather stations became an alternative, which is a more reliable solution. Table 1 shows all the weather stations in ORTDM, their spatial locations and type categorised into automatic and manual weather stations. Figure 1 shows the map for ORTDM, the local municipalities and the weather stations, both automatic and manual weather stations.

Rainfall data were used to compute average annual precipitation for the period 1998 to 2018, determining periods and areas with below normal and above normal average precipitation. Standardised precipitation index (SPI) was used to compute drought frequency, severity and intensity to expose the high drought risk areas. Various indices have been developed to assess the onset, severity, frequency, intensity and end of droughts (Mahlalela, Blamey & Reason 2018). The selection and application of these methods are based on the anticipated objectives, nature of the indicator, local conditions, data availability and data validity (Gerwin et al. 2018; Gqwede 2018; Maza et al. 2019). This study employed the SPI because of its popularity and ability to synthesise long-term data records of precipitation. This study’s rainfall data spanned over 20 years. A number of studies employed the SPI and commended the index (Gqwede 2018; Sprecher 2017). In addition to being widely recommended, this study’s choice of SPI was also influenced by the nature of the available data.

In order for the results to be precise, understandable and presentable, the Meteorological Drought Monitor (MDM) software programme was used to compute the SPI values for the moving average at 3 months, 6 months and 12 months (yearly) for all the stations within ORTDM. Simple imputation method was employed to replace missing precipitation data for stations where missing data were less than 5% and observed data of neighbouring stations or reference stations were used to replace missing data where more than 5% of data were missing (Aieb et al. 2019). The yearly SPI was presented graphically to show the months which ORTDM was vulnerable to droughts throughout the referenced years of study (1998–2018). The MDM output results generated the frequency, intensity of droughts, drought durations, including minimum and maximum drought time lags. Meteorological Drought Monitor software was used to compute both the 3 months SPI moving average (3-SPI) and 6 months SPI moving average (6-SPI).

Results and findings were presented graphically using Microsoft Excel, and Microsoft Word was used to draw tables that presented the results in order to compare the outcomes across ORTDM. One-way analysis of variance (ANOVA) was used to compare SPI values and MDM output for all the local municipalities in ORTDM to uncover the ones that experienced most drought occurrences, level of severity, frequency and intensity and highlight the most vulnerable areas within ORTDM.

Negative SPI values represent rainfall deficit, whereas positive SPI values indicate rainfall surplus. The intensity of drought was classified according to the magnitude of negative SPI values such that the larger the negative SPI values were, the more serious the event was (Otkin et al. 2019). Table 2 below is an SPI interpretation and shows the level of wetness and dryness in the rainfall data of ORTDM. The findings and results of MDM output were interpreted using the SPI Table and these conformed to the SPI interpretation table.

This article followed all ethical standards for research without direct contact with human or animal subjects.

Figure 2A – E below depicts the average monthly precipitation for all the five local municipalities in ORTDM. The KSD municipality received its highest rainfall in summer during the months December, January and February with an average monthly precipitation of 85.6 mm. The average rainfall in summer was higher than the monthly average precipitation of 70.2 mm reported by Mditshwa and Hendrickse (2017) for the same municipality. June is reportedly the driest month in the KSD municipality and the presented results concur with the previous finding by Mahlalela et al. (2018). The KSD municipality has an annual average precipitation 68.8 mm and this is 1.1 mm less than the annual average precipitation of ORTDM.

Figure 2B depicts the average monthly precipitation patterns of Ngquza Hill for the same period (1998–2018). Ngquza Hill also receives most rainfall in December, January and February, receives the least amount of precipitation during winter (May – August) and has a monthly average precipitation of 132.7 mm and an average annual precipitation of 76.2 mm. Coastal areas along Indian Ocean are influenced by warm Mozambique current and as a result winters are wetter and warmer than the areas inland (ed. Thill 2019). Ngquza Hill is a coastal municipality and it is the second of all the local municipalities in the ORTDM that received the highest rainfall. Its average annual and average monthly precipitation was higher than the other three local municipalities in ORTDM except Port St Johns municipality. The average monthly precipitation for Port St Johns municipality is 88.0 mm and 89.9 mm is the annual average precipitation. Port St Johns municipality winters are not as dry as the rest of the other local municipalities in ORTDM and it receives sufficient rainfall and was not susceptible to hydrological drought. Nyandeni Local Municipality and Mhlontlo Local Municipality received an average monthly precipitation of 65.1 mm and 64.1 mm and average annual precipitation of 64.9 mm and 66.5 mm, respectively. They both receive extremely lower rainfall in winter. Indices used to monitor monthly trends of climate in the Eastern Cape detected that areas inland are drier and colder than coastal areas (Mahlalela et al. 2018). Mhlontlo local municipality received reduced precipitation in winter season. Mhlontlo municipality experienced its driest period in 2014 and the authorities there reported loss of livestock and reduced yields on crop production (Wambua 2019).

Agricultural practices in ORTDM are a primary human activity; therefore, it is important to compute the 3-month SPI to reflect short- and medium-term moisture conditions that are a basis to agricultural droughts. The 3 months SPI helped to detect soil moisture, groundwater and reservoir storage (Cammalleri et al. 2019). When SPI is computed for shorter accumulation periods, in this case a period of 3 months, it is used as an indicator of reduced soil moisture and this has an important impact on agriculture and crops, especially during farming seasons. Figure 3A – E below shows the graphical representation of the 3 months SPI values distribution for all the weather stations in all the five local municipalities in the ORTDM for the period 1998–2018.

Figure 3A – E illustrates that all the five municipalities are susceptible to droughts of varying magnitude and frequency and the variation is noticeable in specific areas serviced by different weather stations. The presented graphs also show incidence of extreme to severe droughts of negative 3 months SPI values above −2.5. The incidence of meteorological, hydrological and agricultural droughts in the district municipality is not peculiar. Despite high negative drought severity figures across all the weather stations, the presented bell curves reflect higher proportion of near normal and moderately wet conditions. This is the case for all areas as per their respective weather stations. Altın, Sarış and ltın (2019) and SAWS (2018) reported similar drought patterns indicating that some parts of the country experience below-normal rainfall at varying frequencies with some areas being drier than others during the same period. These results concur with a number of previous studies in the Eastern Cape and South Africa in general (Altin et al. 2019). Mhlontlo Local Municipality is one of the driest inland municipalities in ORTDM and experienced multiple droughts negatively impacting agriculture and water sources between 1998 and 2018 (Dotse 2019). Overall, ORTDM is susceptible to droughts of varying intensity and the results of the 3 months SPI across the entire district municipality confirm this. The SPI drought categories from 1998 to 2018 of ORTDM show that some of the levels of hydrological droughts experienced have the potential to cause some devastating impact such as shortage of drinking water and reduced crop yields. Droughts negatively impact river flows, dam levels, crop yields and animal life (Jimmy et al. 2019). The presented spatial analyses of drought severity, recurrence nature of droughts and the level of variance across different areas could be used as a tool for identification of the most drought prone areas and drought periods and these could assist in resource allocation for drought preparedness.

Weather stations in the same municipalities show significant differences in SPI values. This, therefore, implies that the quality of drought data and information could be improved by increasing the density of weather stations in an area. The KSDM has the highest number of weather stations and the difference in SPI values across KSD municipality shows that the same municipality can experience varying degrees of drought in one period and this should not be generalised per municipality. A similar conclusion was made by Lucinda and Baez-Villaneuva (2019). When conducting a drought-related study, it is very important to look into the weather stations and compare the findings rather than generalising. Some critically affected areas might be overlooked because of a collective description of areas and results (Forbes & Cyr 2019).

The following section quantifies the amount of the time each of the five local municipalities in the ORTDM experienced drought and puts them in different categories using both 3 and 6 months SPI values. Tables 3 and 4 show 3 and 6 months SPI results for all the five local municipalities, respectively. The SPI results have been tabulated according to drought severity class of moderately dry, severely dry and extremely dry categories and computed for each municipality.

The summarised results in Table 3 show the summarised results per municipality. The results show that all the local municipalities in ORTDM experienced droughts of varying severity at different time proportions. The KSD Local Municipality has the highest probability of experiencing a drought with a percentage of 61.2%, followed by Mhlontlo, Ngquza Hill and lastly PSJ and Nyandeni with equal probabilities of 37%. Using the 3 months SPI values and summing up all the drought periods when the SPI is below −1, ORTDM has a 56.5% chance of experiencing drought. Agriculture is the main economic activity in ORTDM and October, November and December are the growing months. A more than 50% chance of drought in area has important implications on livelihoods and food security in areas where agriculture is the main livelihood. The higher incidence of droughts in OR Tambo was also reported by Bae et al. (2019).

In addition to results on 3 months SPI drought severity classes presented in Table 3, Table 4 presents results of the 6 months SPI for the same municipality. Unlike the 3 months SPI results which estimated a probability of 56.5% susceptibility to droughts for ORTDM, the 6 months SPI estimated a probability of 62%. Thus, the district is likely to experience more cumulative 6 months’ droughts than 3 months’ droughts. These results are in line with a study by Jimmy et al. (2019) who noted that OR Tambo DM experiences hydrological droughts that intensify into agricultural droughts in a period of 6 months. These results also confirmed reports documented by the Department of Water Affairs and Forestry (DWAF) on negative anomalies perceived in ORTDM for both on surface and subsurface water (Mditshwa & Hendrickse 2017). A report by Bae et al. (2019) highlighted only six agricultural droughts that occurred in ORTDM and these occurred in 2013, 2014, 2015, 2016 and 2017 (Bae et al. 2019). However, none of the previous studies presented drought reports according to drought severity class of moderately dry, severely dry and extremely dry categories. It is important to note that literature has vast information on droughts in the Eastern Cape and ORTDM as shown by multiple citations above but the existing information lacks some level of standardisation that can warrant comparability across both space and time. This owes to the fact that drought studies lack standardisation in terms of methodology, classification and scales used, a challenge also mentioned by Masupha and Moeletsi (2017). Widespread adoption and SPI or any methodology that promotes the calculation of the probability of droughts in any area could not be overemphasised, especially, for drought preparedness purposes.

Standardised precipitation index data and analyses characterise droughts in a way that detect both the onset and cessation of drought incidents, something that other indices are unable to do (Cook 2019). It also determines drought severity, frequency and intensity of droughts and identifies areas of high vulnerability. Information on average drought intensity and duration is crucial for decision-making purposes. Table 5 presents minimum and maximum drought intensity, average drought intensity and maximum drought duration for all the five local municipalities in ORTDM.

Average drought intensity ranges from −0.62 to −0.99. Ngquza Hill Local Municipality is a coastal municipality. The 3 months SPI values for Ngquza Hill are in contrast to the known supposition and theory that coastal areas are wetter than inland areas (Sotsha 2013). The SPI detects the onset and end of the drought. Despite presenting a mean drought intensity that is close to normal, all the local municipalities experience long periods of drought ranging from 30 to 80 days of consecutive dry days. Mhlontlo local municipality experienced the longest drought duration of 80 days and all the municipalities experienced their longest drought duration in 2014. The results of the 2014 SPI further confirm the results by Mafongoya et al. (2019) who assert that in 2014, the whole African continent experienced its worst drought in more than 50 years with life-threatening and devastating impacts. The two coastal municipalities, PSJ and Ngquza Hill, have relatively lower average drought duration. The presented outcome confirms that Nyandeni is the driest municipality as it experiences the maximum intensity and lengthy drought duration. The results agree with the previous studies which asserted that areas inland are drier than coastal areas as noted by Mackay and Gross (2019) in drought studies conducted in Australia. In South Africa, the years 2014 and 2015 were drought years and Port St Johns Local Municipality was amongst local municipalities in ORTDM that experienced water stress and reported a reduction in water levels in its water bodies and reduced crop yields in farms (Mantsho 2018).

The SPI results could provide significant statistics that could be considered for drought monitoring, drought resources allocation and drought preparedness. When distributing drought relief aid, government departments and aid agencies should prioritise inland municipalities. In addition, these results are crucial for agricultural purpose. Accordingly, Nyandeni and Mhlontlo should be highly prioritised when it comes to agricultural drought intervention strategies such as introduction of drought tolerant plants and animals. This should be the case, especially when agriculture is a key livelihood activity. The maximum drought duration for Mhlontlo, Nyandeni and KSD in ORTDM shows that these municipalities are vulnerable and high-risk municipalities and the same was echoed by Dotse (2016). The findings of this study confirmed the hypothesis of this study that ORTDM is susceptible to hydrological droughts and furthermore, revealed the extent of drought effect, frequency, level of severity and intensity and detailed the areas of higher vulnerability.